[RFA] symtab.c: Search section table when fixing up a symbol'ssection

[Kevin Buettner <kevinb at redhat dot com>]

The patch below fixes all scope.exp failures for the frv-uclinux target.

I believe that fixup_section() was broken for many other targets too,
but innocuously so, since most targets relocate all sections (for a
particular objfile) by a single constant.  Thus, for those targets, it
doesn't matter that fixup_section() incorrectly computes (or fails to
compute) a symbol's section index.  OTOH, the frv-uclinux target
requires that the section index be accurately computed since read-only
and read/write sections are relocated by different amounts.  Other
targets with this property (such as AIX) avoid the problem by making
sure that the section value is set correctly by the symbol reader itself.

Other details are included in a comment in the patch itself.  (See
below.)

Thanks to Jim Blandy for suggesting the section table search.  My
first attempt at fixing this problem used a much less efficient
search of the minimal symbols.

Okay to commit?

	* symtab.c (fixup_section): Search section table when lookup by
	name fails.

Index: symtab.c
===================================================================
RCS file: /cvs/src/src/gdb/symtab.c,v
retrieving revision 1.129
diff -u -p -r1.129 symtab.c
--- symtab.c	8 Apr 2004 21:18:13 -0000	1.129
+++ symtab.c	17 May 2004 18:57:34 -0000
@@ -868,6 +868,62 @@ fixup_section (struct general_symbol_inf
       ginfo->bfd_section = SYMBOL_BFD_SECTION (msym);
       ginfo->section = SYMBOL_SECTION (msym);
     }
+  else if (objfile)
+    {
+      /* Static, function-local variables do appear in the linker
+	 (minimal) symbols, but are frequently given names that won't
+	 be found via lookup_minimal_symbol().  E.g., it has been
+	 observed in frv-uclinux (ELF) executables that a static,
+	 function-local variable named "foo" might appear in the
+	 linker symbols as "foo.6" or "foo.3".  Thus, there is no
+	 point in attempting to extend the lookup-by-name mechanism to
+	 handle this case due to the fact that there can be multiple
+	 names.
+	 
+	 So, instead, search the section table when lookup by name has
+	 failed.  The ``addr'' and ``endaddr'' fields may have already
+	 been relocated.  If so, the relocation offset (i.e. the
+	 ANOFFSET value) needs to be subtracted from these values when
+	 performing the comparison.  We unconditionally subtract it,
+	 because, when no relocation has been performed, the ANOFFSET
+	 value will simply be zero.
+	 
+	 The address of the symbol whose section we're fixing up HAS NOT
+	 BEEN adjusted (relocated) yet.  It can't have been since the
+	 section isn't yet known and knowing the section is necessary
+	 in order to add the correct relocation value.  In other
+	 words, we wouldn't even be in this function (attempting to
+	 compute the section) if it were already known.
+
+	 Note that it is possible to search the minimal symbols
+	 (subtracting the relocation value if necessary) to find the
+	 matching minimal symbol, but this is overkill and much less
+	 efficient.  It is not necessary to find the matching minimal
+	 symbol, only its section.  
+	 
+	 Note that this technique (of doing a section table search)
+	 can fail when unrelocated section addresses overlap.  For
+	 this reason, we still attempt a lookup by name prior to doing
+	 a search of the section table.  */
+	 
+      CORE_ADDR addr;
+      struct obj_section *s;
+
+      addr = ginfo->value.address;
+
+      ALL_OBJFILE_OSECTIONS (objfile, s)
+	{
+	  int idx = s->the_bfd_section->index;
+	  CORE_ADDR offset = ANOFFSET (objfile->section_offsets, idx);
+
+	  if (s->addr - offset <= addr && addr <= s->endaddr - offset)
+	    {
+	      ginfo->bfd_section = s->the_bfd_section;
+	      ginfo->section = idx;
+	      return;
+	    }
+	}
+    }
 }
 
 struct symbol *